在AI中用Cavalry创建。一个MCP服务器，为你的AI助手提供直接进入Cavalry的通道——运行脚本，查询知识库，并根据提示进行动画制作。
概览
- — Claude 通过 Stallion 桥实时执行 Cavalry JavaScript - — 语义搜索 40,000+ 块的骑兵文档、真实场景和社区知识 - — 读取并保存你的活跃场景 - — 自动加载每次会话经过验证的最佳实践参考 从GitHub Releases下载lancedb_public.zip，提取到项目中： LanceDB 需要本地硬盘。设置绝对本地路径：LANCEDB_PATH=C：/Users/yourname/.cavalry-assistant/lancedb_public Stallion 脚本>——保持它运行在8080端口。 该文件夹中的.mcp.json在用Claude代码打开时会自动注册。或者手动： 公开知识库包括Discord社区数据（私密）。 - Stallion 总是响应“成功”——请查看 Cavalry 控制台面板是否有实际错误- 始终用 api.stop（） 启动场景修改脚本 - 完整 API 参考：提示符/骑兵最佳实践。
概述（AI 翻译）
WRA 骑兵助理
Claude 实时编写并运行 Cavalry 脚本——不复制粘贴，不切换窗口。
它的作用
- 写入并运行脚本 — Claude 通过 Stallion 桥实时执行 Cavalry JavaScript
- 懂文档 — 语义上搜索40,000+块骑兵文档、真实场景和社区知识
- 场景感知 — 读取并保存你的活动场景
- 保持最新 — 每次会话自动加载经过验证的最佳实践参考
要求
| 工具 | 目的 |
|---|---|
| 骑兵 | 应用 — 需要启用种马 |
| Node.js 18+ | 运行MCP服务器 |
| Ollama | 局部嵌入 |
| Claude 代码 | 人工智能客户端 |
快速设置（预建知识库）
1.克隆与安装
2.下载知识库
从 GitHub Releases 下载 lancedb_public.zip，提取到项目中：
更新 .env：
Windows / 网络驱动器用户： LanceDB 需要本地驱动器。设置绝对本地路径：
LANCEDB_PATH=C:/Users/yourname/.cavalry-assistant/lancedb_public
3.开始，奥拉玛
4.骑兵中的种马
本页由 AI 翻译，技术细节请以原文为准。
README
以下为原文 README 节选，完整内容请查看原文。
What it does
- Write & run scripts — Claude executes Cavalry JavaScript live via the Stallion bridge
- Knows the docs — searches 40,000+ chunks of Cavalry documentation, real scenes, and community knowledge semantically
- Scene-aware — reads and saves your active scene
- Stays current — auto-loads a verified best-practices reference on every session
Requirements
| Tool | Purpose |
|---|---|
| Cavalry | The app — needs Stallion enabled |
| Node.js 18+ | Runs the MCP server |
| Ollama | Local embeddings |
| Claude Code | The AI client |
Quick Setup (pre-built knowledge base)
1. Clone and install
git clone cavalry-assistant
cd cavalry-assistant
cp .env.example .env
npm install --prefix mcp
npm run build --prefix mcp
2. Download the knowledge base
Download lancedb_public.zip from GitHub Releases, extract into the project:
cavalry-assistant/
data/
lancedb_public/ ← extract here
Update .env:
LANCEDB_PATH=./data/lancedb_public
Windows / network drive users: LanceDB requires a local drive. Set an absolute local path:
LANCEDB_PATH=C:/Users/yourname/.cavalry-assistant/lancedb_public
3. Start Ollama
ollama pull nomic-embed-text
4. Enable Stallion in Cavalry
Scripts > Stallion — leave it running on port 8080.
5. Connect to Claude Code
The .mcp.json in this folder auto-registers when opened in Claude Code. Or manually:
claude mcp add cavalry-assistant node mcp/dist/index.js
6. Start a session
/cavalry
Knowledge Base
| Source | Content |
|---|---|
| Official docs | Full Cavalry documentation |
| Scenery scenes | 270 real .cv scene files — real-world node/connection patterns |
| Manual reference | Curated scripting patterns and best practices |
| Scripts | Example JavaScript scripts |
The public knowledge base does not include Discord community data (private).
Building the Knowledge Base (contributors)
1. Set up Python
cd etl
pip install -r requirements.txt
2. Configure .env
LANCEDB_PATH=C:/Users/yourname/.cavalry-assistant/lancedb
DISCORD_TOKEN=your_token
DISCORD_CHANNEL_IDS=channel_id_1,channel_id_2
SCENERY_COOKIE=your_session_cookie
3. Run ETL
cd .. # project root
# Public sources
python etl/ingest.py --source docs
python etl/ingest.py --source scenery
python etl/ingest.py --source manual
python etl/ingest.py --source scripts安装
This server does not publish a one-line install command.
Open the repository installation guide 新手？先看 mcpskill.org 教程 →